DESCRIPTION (Applicant's Abstract): The long-term objective of this proposal is to enhance the therapeutic efficacy of chemotherapeutic regimens by understanding and ultimately manipulating the expression of CpG-island-containing genes. CpG islands are small, CpG-enriched DNA sequences associated with the promoters of approx. 50% of all human genes. In normal cells, they are devoid of cytosine methylation and contain within their nuclease-hypersensitive regions that bind transcription factor (TF) and stimulate transcription. In tumors, CpG island-containing genes can be transcriptionally silenced by a mechanism involving CpG island methylation and loss of TF accessibility. It may be possible to trigger this silencing process in a potentially therapeutic manner. Development of such gene silencing strategies is hindered, however, by a relative lack of knowledge about how CpG island function, and specifically about which CpG islands certainly play a role in gene expression. The applicant's recent studies suggested that non- transcription factor binding regions also contribute to CpG island- containing gene expression, and do so in a methylation-dependent manner. This hypothesis will be tested using, as a model, the CpG island of the human MGMT gene. The specific aims are: (1) to determine if the non-TF binding regions of CpG islands contribute to gene expression; (2) to determine if cytosine methylation blocks the ability of non-TF-binding regions of CpG islands to gene expression; and (3) to determine if the ability of non-TF-binding regions of CpG islands to contribute to methylation-dependent gene expression is a generalized phenomenon. These studies will contribute to the understanding of CpG island function and represent a first step in the potential manipulation of the many therapeutically relevant genes (MGMT, GST-pi, p53, and mdr1) whose expression is controlled by CpG islands.